Nano Res
1
Benchtop aqueous two-phase extraction of isolated
individual single-walled carbon nanotubes.
Navaneetha K. Subbaiyan,1,† A. Nicholas G. Parra-Vasquez, 1,† Sofie Cambré,1,2,† Miguel A.
Santiago Cordoba,1 Sibel Ebru Yalcin,1 Christopher E. Hamilton,1 Nathan H. Mack,1 Jeffrey L.
Blackburn,3 Stephen K. Doorn,1 and Juan G. Duque1().
Nano Res., Just Accepted Manuscript • DOI: 10.1007/s12274-014-0680-z
http://www.thenanoresearch.com on December 8 2014
© Tsinghua University Press 2014
Just Accepted
This is a “Just Accepted” manuscript, which has been examined by the peer-review process and has been
accepted for publication. A “Just Accepted” manuscript is published online shortly after its acceptance,
which is prior to technical editing and formatting and author proofing. Tsinghua University Press (TUP)
provides “Just Accepted” as an optional and free service which allows authors to make their results available
to the research community as soon as possible after acceptance. After a manuscript has been technically
edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP
article. Please note that technical editing may introduce minor changes to the manuscript text and/or
graphics which may affect the content, and all legal disclaimers that apply to the journal pertain. In no event
shall TUP be held responsible for errors or consequences arising from the use of any information contained
in these “Just Accepted” manuscripts. To cite this manuscript please use its Digital Object Identifier (DOI®),
which is identical for all formats of publication.
Nano Research
DOI 10.1007/s12274-014-0680-z
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
1 Nano Res.
Template for Preparation of Manuscripts for Nano
Research
This template is to be used for preparing manuscripts for submission to Nano Research. Use of this
template will save time in the review and production processes and will expedite publication.
However, use of the template is not a requirement of submission. Do not modify the template in
any way (delete spaces, modify font size/line height, etc.). If you need more detailed information
about the preparation and submission of a manuscript to Nano Research, please see the latest
version of the Instructions for Authors at http://www.thenanoresearch.com/.
TABLE OF CONTENTS (TOC)
Benchtop aqueous
two-phase
extraction of
isolated
individual
single-walled
carbon nanotubes
Navaneetha K.
Subbaiyan, A.
Nicholas G.
Parra-Vasquez,
Sofie Cambré,
Miguel A.
Santiago Cordoba,
Sibel Ebru Yalcin,
Christopher E.
Hamilton, Nathan
H. Mack, Jeffrey
L. Blackburn,
Stephen K. Doorn,
and Juan G.
Duque*
Los Alamos
National
Laboratory, USA
University of
Antwerp, Belgium
National
Renewable Energy
Laboratory, USA
A practical, fast and scalable benchtop aqueous-two-phase extraction of isolated
individual single-walled carbon nanotubes from bundles and impurities in aqueous
dispersions is presented.
Provide the authors’ website if possible.
http://www.lanl.gov/expertise/profiles/view/juan-duqe
Top Phase
Bottom Phase
Impurities
Individuals
300 600 900 1200 15000
2
4
6
8
10
300 600 900 1200 1500
Ab
so
rptio
n
Before separation
Isolated
Bundles
Wavelength (nm)
Benchtop aqueous two-phase extraction of isolated
individual single-walled carbon nanotubes.
Navaneetha K. Subbaiyan,1,† A. Nicholas G. Parra-Vasquez, 1,† Sofie Cambré,1,2,† Miguel A. Santiago
Cordoba,1 Sibel Ebru Yalcin,1 Christopher E. Hamilton,1 Nathan H. Mack,1 Jeffrey L. Blackburn,3 Stephen K.
Doorn,1 and Juan G. Duque1().
Received: day month year
Revised: day month year
Accepted: day month year
(automatically inserted by
the publisher)
© Tsinghua University Press
and Springer-Verlag Berlin
Heidelberg 2014
KEYWORDS
Carbon nanotubes,
aqueous two-phase
separation (ATP),
aggregate removal,
isolation, sorting
ABSTRACT
Isolation and purification of single-walled carbon nanotubes (SWCNTs) is a
prerequisite for their implementation in various applications. In this work, we
present a fast (~5 minutes), low cost and easily scalable benchtop approach to
extract high-quality isolated SWCNTs from bundles and impurities in an
aqueous dispersion. The extraction procedure, based on aqueous two-phase
separation (ATP), is widely applicable to any SWCNT source (tested up to
diameters of 1.7nm), and independent of defect density, purity, diameter, and
length. The extracted dispersions show comparable removal of large
aggregates, small bundles and impurities to density gradient
ultracentrifugation, but without the need for any high end instrumentation.
Raman and fluorescence-excitation spectroscopy, single-nanotube fluorescence
imaging, atomic force and transmission electron microscopy, and
thermo-gravimetric analysis all confirm the high purity of the isolated
SWCNTs. By predispersing the SWCNTs without sonication (only gentle
stirring) full-length, pristine SWCNTs can be isolated (tested up to 20µm).
Hence, this simple ATP method will find immediate application in generating
SWCNT materials for all levels of nanotube research and applications, from
fundamental studies to high performance devices.
1. Introduction
Single-walled carbon nanotubes (SWCNTs) have
been idealized for their intrinsic properties, which
depend on their chiral structure [1]. However, purity,
heterogeneity, and processability of the
as-synthesized materials have limited their
widespread implementation in electronics, photonics,
and sensor applications. To obtain high purity,
isolated SWCNTs in solution, suspension agents have
Research Article
Nano Research
DOI (automatically inserted by the publisher)
Address correspondence to Juan G. Duque, [email protected]
| www.editorialmanager.com/nare/default.asp
2 Nano Res.
been used to overcome van der Waals forces between
SWCNTs [2-6] combined with strong ultrasonication
to break apart bundles [2, 3, 5], inadvertently cutting
the SWCNTs in short segments [7].
After predispersion, individual SWCNTs remain in
solution along with small carbonaceous and metallic
particles and bundles 8, 9], which perturb SWCNT
band structure as evidenced by changes in optical
signatures [10, 11]. In particular, bundling of
SWCNTs can cause red-shifting and broadening in
absorption, quenching of fluorescence, and shifting
of vibrational Raman signatures [11, 12]. Different
techniques have been developed to isolate individual
SWCNTs from bundles and other impurities [2-5, 9,
13, 14], with most of these techniques depending on
an initial ultracentrifugation (UCF) step followed by
a more dedicated technique such as density gradient
ultracentrifugation (DGU) [9, 13]. Quantitative
techniques, such as rheology and atomic force
microscopy (AFM), estimate that after UCF up to 50%
of bundles remain [5, 8, 15] and further purification
by DGU results in only 5-15% of bundles still present
[15, 16].
Both UCF and DGU rely on sedimentation to
separate individual SWCNTs from other impurities.
UCF is based on sedimentation speed where more
dense materials travel more rapidly to the bottom. In
contrast, DGU relies upon equilibrium-based
sedimentation, where surfactant organization around
SWCNTs of different structure and electronic type
define the density of dispersed objects that then
sediment to a level of equal density [9, 13, 15, 17].
Arnold et al. [15] determined with analytical
centrifugation that the sedimentation coefficients of
bundles and isolated SWCNTs strongly overlap,
making their complete separation by UCF unlikely,
while DGU can more efficiently remove bundles.
Despite the DGU sorting capability, the quality
depends on the initial isolation of the predispersion,
where higher concentration of bundles will blur the
separation lines [9]. All techniques offering increased
purity, such as DGU, entail higher costs (including, in
particular, use of expensive ultracentrifuges), longer
times, more complexities, multiple steps, and lower
yields. Thus to streamline commercial advancement,
a quick, scalable, low-cost, and high-yield technique
is needed. Such a low-cost purification method to
remove SWCNT bundles and residual metallic
catalyst was proposed with the use of permanent
magnets instead of high-end instrumentation [14].
However, since the method relies on differences in
magnetic sedimentation of bundles and isolated
SWCNTs, compromise must be made on separation
time (several hours) versus purity, never able to reach
the separation quality of DGU.
Recently, aqueous two-phase separation (ATP) has
been introduced as a highly scalable and fast method
to separate a wide range of chirality-enriched
SWCNT materials [18-22]. In this work we show that
the ATP method can also be adjusted to extract
isolated SWCNTs and to eliminate production
impurities (aggregates, carbonaceous impurities and
SWCNT bundles), irrespective of the source materials
(HiPco, CoMoCat, Plasma Torch, Arc Discharge and
Laser Vaporization). Moreover, our method is
extremely fast (~5 min), can simultaneously yield
isolation of single chiralities, and is independent of
length, achieving DGU-quality isolated individual
SWCNTs either from solely stirred (i.e. long,
undamaged SWCNTs, lengths up to 20 μm) or
sonicated dispersions (i.e. short SWCNTs, <1 μm).
The benchtop procedure described does not require
expensive and specialized equipment (like an
ultracentrifuge), thereby facilitating increased
widespread availability of high quality (full-length)
isolated SWCNTs.
2. Experimental
SWCNT materials were obtained from various
synthesis sources. Table S1 in the Electronic
Supplementary Material (ESM) provides an overview
of the synthesis source, batch number, and SWCNT
diameter range (obtained from UV-vis-NIR
absorption). Sodium deoxycholate (DOC, AMRESCO
Lot # 0331C075), polyethyleneglycol (PEG, mol.wt.
6000 Da, Alfa Aesar Lot# 10173268), and dextran for 3
different mol. wt. (DX, mol.wt. 70000 Da, TCI
Chemicals, Lot # NSBSF-TS; mol.wt. 9000-11000 Da,
Sigma Aldrich, Lot # BCBK8656V; and mol.wt.
35000-45000 Da, Sigma Aldrich, Lot # SLBD3835V)
were used as received.
SWCNT materials were dispersed in a 1.04%wt/V
DOC aqueous solution, either by sonication
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
3 Nano Res.
(typically 1 hour tip sonication, Sonic Vibra Cell with
tip CV18-9909, operated at 8W) or by gentle stirring
over a period of 1-3 weeks. For the sonicated
suspensions, 1mg/mL raw SWCNT material was
dispersed, while for the stirred suspensions a slightly
higher starting concentration (1.6mg/mL) was used
to achieve better isolation [4].
For separation (unless otherwise stated), these
dispersed SWCNTs were added to a PEG:dextran
(9%wt/V : 9%wt/V in H2O, with mol.wt. of dextran
70000Da) stock solution, such that a final DOC
concentration of 0.088%wt/V was achieved. A typical
separation consisted of adding 400µL of the parent
suspension (1.04%wt/V DOC) and 240µL of H2O to
4080µL of the 9:9%wt/V PEG:dextran solution. The
PEG:dextran stock solution was mixed overnight to
dissolve the PEG and dextran in water. Immediately
after formation of the mixture, the
SWCNT/PEG:dextran suspension was benchtop
centrifuged for 5 minutes at 6000g to achieve phase
separation, and top and bottom fractions were
collected manually by a syringe. Unless otherwise
stated, all separations were performed at room
temperature (22±2°C). Separations for larger
diameter SWCNT samples were performed similarly,
using D2O instead of H2O, which was preferred
because of its higher optical transparency in the NIR.
Ultracentrifugation (UCF) and centrifugation (CF) of
the pre-dispersed parent suspension, only used for
control experiments, were performed on a
Thermo-Scientific WX-80 Ultra series ultracentrifuge.
For CF, 2 hours at 40000g (17000rpm, TH-641
swing-out rotor) was applied. For UCF comparison, 1
hour at 287000g (40000rpm) was used. For DGU
control experiments, a gradient was prepared inside
thick-wall polyalomer 2.5mL centrifuge tubes, using
iodixanol as gradient medium (Sigma Aldrich, D1556,
60%wt/V). A lower density layer (5% iodixanol and
1%wt/V DOC, 1.4mL), containing the pre-dispersed
SWCNTs, was superimposed on a high density layer
(27% iodixanol and 1%wt/V DOC, 1mL). The
centrifuge tube was tilted (angle ~40°) and rotated
around its axis to obtain a quasi-linear gradient.
Centrifugation was performed for 20h at 270000g
(45000rpm; SW60-Ti swing-out rotor) and the
resulting isolated SWCNTs were selected manually
by a syringe.
Dialysis of the separated SWCNT phases was
performed to remove the polymers after phase
separation, using a pressurized Millipore Stirred
Ultrafiltration Cell equipped with regenerated
cellulose membranes (100kDa). Each phase was
filtered until 10% of the volume remained and then
1%wt/V DOC solution was added to dilute the
remnants by a factor of 10. This was repeated several
times to achieve a minimum dilution of 105 for PEG
and dextran. Absorption and resonant Raman spectra
were acquired to test that this did not result in
quantifiable new bundle formation.
For a description of the different experimental
techniques, variations of the ATP sorting conditions
and data treatments we refer to the ESM.
3. Results and Discussion
In ATP separations, two water-soluble but immiscible
polymers, e.g. PEG and dextran, are mixed together
after which they spontaneously form two phases
with different hydrophobicity [23]. The resulting top
phase, i.e. PEG-rich phase, is less hydrophilic than
the dextran-rich bottom phase. Tuning the surfactant
structure around the SWCNT surface results in
differences in relative hydrophobicities of various
SWCNT species [20]. Different chiralities thus
separate into the different phases. With our current
understanding of the separation mechanism [20], we
sought to develop a technique to separate isolated
SWCNTs from bundles and other production
impurities.
We started from a stock dispersion of PEG and
dextran, both with 9% wt/V concentration, and
added different amounts of a 1.04%wt/V DOC -
SWCNT parent suspension to this stock solution
(with corresponding H2O addition to have similar
dilution factors), and subsequently induced the
phase-separation by benchtop centrifugation for 5
minutes at 6000g. Figure 1a presents the absorption
spectra of a sonicated parent HiPco-SWCNT
dispersion (in black), and spectra of top (in blue),
bottom (in red), and interface (in green) after phase
separation with a final DOC concentration of 0.088%
wt/V. Unless otherwise stated, for proper comparison,
absorption spectra of top, bottom, and interface are
rescaled, according to their specific volumetric
Figure 1: Absorption spectra of ATP separated HiPco 195.2 SWCNTs. (a) Absorption spectra of the parent HiPco 195.2 suspension
(black), bottom (red), top (blue) and interface (green) for an ATP separation with a final DOC concentration of 0.088%wt/V. The bottom
(1180µL) and top (3540µL) absorption spectra were rescaled according to their volumetric dilution (see ESM), while the interface
(unknown volume) was rescaled so that SUM=bottom + top + interface (shown in magenta) matches the absorption signal of the parent
suspension. The inset shows an expanded absorption spectrum of the ATP bottom separated phase, in comparison to a rescaled
absorption spectrum of the top (divided by 7) and parent (divided by 20) suspensions, revealing a much higher SWCNT - to -
background ratio in the absorption spectra of the bottom phase. Samples were diluted if necessary for the absorption measurements
(Optical Density < 3) and afterwards rescaled according to this dilution factor. (b) Absorption spectra of the ATP bottom phases for
different final DOC concentrations, showing that 0.088%wt/V ideally removes the impurity-related background. (c) Comparison of the
absorption spectra of the parent HiPco 195.2 SWCNT dispersion (black) and bundle-removal by CF (2h @ 40.000g, green), UCF (1h @
287.000g, cyan), DGU ([email protected], blue) and ATP (5 min @ 6.000g, red). Parent, CF and UCF spectra are presented as measured,
ATP was corrected for its volumetric dilution and DGU was rescaled on the intensity of the excitonic transitions of the ATP separation,
showing that the background - to - SWCNT absorption is identical for ATP and DGU, while in CF and UCF a significant fraction of the
background remains present. (d) Absorption spectra of the ATP separated (red), DGU (blue), UCF (cyan) and parent SWCNT dispersion
(black), diluted and rescaled to obtain a normalized PL intensity after excitation at 570nm (see inset for comparison of parent and ATP
separated PL, DGU and UCF yield similar PL spectra but are omitted for clarity). All absorption data are cut around 1400nm due to
strong H2O absorbance in this region.
dilution after phase separation (for details see ESM
section S2.a), as previously reported [20]. After
proper rescaling, the sum of bottom, top, and
interface (in magenta) matches the absorption
signatures of the parent suspension.
The absorption spectrum of the bottom phase (red
curve in Fig. 1a) shows sharp excitonic transitions of
all isolated SWCNT chiralities, but with a strong
reduction in background (in comparison with the
parent suspension). In contrast, the more
hydrophobic top phase (and the interface) contains a
strong absorption 1/λ-like scattering background
(arising from all scatterers, including larger
aggregates) and highly broadened excitonic
transitions (arising from small aggregates:
SWCNT-SWCNT interactions typically broaden and
red-shift SWCNT transitions [11, 12]), suggesting a
300 600 900 1200 15000
2
4
6
8
10
300 600 900 1200 15000.0
0.5
1.0
1.5
2.0
300 600 900 120015000.0
0.5
Ab
s (
1m
m)
Parent
Bottom
Top
Interface
SUM
Wavelength (nm)
300 600 900 1200 15000
2
4
6
Ab
s (
1m
m)
Parent
CF (2h @ 40000g)
UCF (1h @ 287000g)
DGU
ATP
300 600 900 1200 15000.00
0.05
0.10
0.15
Wavelength (nm)
(d)
(b)
(c) Parent
ATP
DGU
UCF
Ab
s (
1m
m)
Wavelength (nm)
ex
=570nm
(a)
900 1000 1100 1200
PL
In
ten
sity
Emission Wavelength (nm)
Final DOC concentration
0.082%
0.088%
0.095%
0.110%
Ab
s (
1m
m)
Wavelength (nm)
PARENT/ 20
TOP / 7
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
5 Nano Res.
significant difference in solution content of both
phases. The strong background removal is also
visualized in the inset of Fig. 1a, where spectra of
parent and top suspensions are rescaled on this
background, showing much smaller contributions of
resonant SWCNT absorption (due to much higher
relative contribution of background in those samples).
We hypothesize that such significant spectroscopic
differences are due to the removal of aggregates,
carbonaceous impurities, and small SWCNT bundles.
Figure 1b presents rescaled absorption spectra of the
bottom phases after varying the final DOC
concentration for the separation, which was achieved
by stepwise varying the amounts of added H2O and
parent suspension (without changing the total
dilution factor). The scaled spectra demonstrate
that a slightly higher DOC concentration (compared
to the one used in Fig. 1a, from 0.088 to 0.110%)
results in the appearance of additional background
absorption, while hardly any relative intensity
increase for the excitonic SWCNT absorbance can be
observed. A slightly lower DOC concentration
(0.082%) than 0.088%, provides no further removal of
background signal, but results in partial removal and
aggregation of specific chiralities (e.g. the (10,2)
chirality with E11=1062nm). A final DOC
concentration of 0.088%wt/V was found to be ideal
for the removal of the large scattering background in
the absorption spectra, without removing specific
chiralities from the bottom phase.
Figure 1c compares this ATP separation (bottom
phase) with separation of isolated SWCNTs starting
from the same parent suspension via normal
centrifugation (CF; 2h at 40000g), UCF (1h at
2870000g), known to retain 15-50% bundles of
smaller than 4nm in diameter [5, 8, 15], and DGU
(20h at 270000g), known to retain 5-15% of bundles [8,
9, 15]. When comparing peak-to-valley ratios for the
(6,5) chirality (i.e. peak = 982nm, valley = 901nm) for
the different separations, we obtain 1.34 for the
parent suspension, 1.8 for CF, 2.24 for UCF, 3.35 for
DGU, and 3.35 for ATP, thus showing significant
background reduction from the parent dispersion for
all methods, but with ATP and DGU achieving the
greatest reduction. Most importantly, the ratios
clearly show identical results for DGU and ATP,
indicating an equivalent removal of bundles and
other impurities by ATP as by DGU.
A further measure of the improved isolation of the
ATP-extracted SWCNTs comes from an increased
sample PL intensity, which is highly sensitive to
environmental effects and dispersion quality [24]. For
accurate comparison, ATP separated bottom fractions
and DGU samples were first dialyzed into 1 %wt/V
DOC to remove PEG and dextran. All samples were
diluted to eliminate self-absorption effects (optical
density (OD) < 0.05 in a 1 mm cell) and such that
their (6,5) PL intensity exciting at 570 nm is identical
(inset Fig. 1d inset and Fig. S1 in ESM). Since the PL
mainly originates from isolated SWCNTs and not
non-fluorescent absorbing impurities, a much higher
concentration, overall absorption, of the parent
dispersion is needed to achieve the same PL intensity
as the purified dispersions. Moreover, the slightly
Figure 2: Resonant Raman spectroscopy (RRS) of the ATP
separated samples. (a) RRS spectra of the RBM-region for HiPco
195.2 excited at 3 different wavelengths, showing the radial
breathing modes of different chiralities, e.g. (9,7), (10,5), (11,3)
and (12,1) SWCNTs in resonance at 785nm (left to right). The
RBM of the (10,2) SWCNT, i.e. 266cm-1 (indicated by the red
dashed line), is clearly resonant at 725nm, while at 785nm
excitation this RBM is only resonant when bundles are present
[12]. Similarly the RBM of the (9,1) SWCNT, indicated by the
black dashed line at 310cm-1 is resonant at 676.4nm, while at
725nm this RBM is only resonant when bundles are present.
Bundles are thus mainly present in the top phase and at the
interface after bundle-extraction. (b) RRS spectra of the D- and
G-bands excited at 514.5nm and 676.4nm, for the HiPco 195.2
and Plasma Torch SWCNTs. D/G ratios are presented in
parenthesis. The ATP bottom fraction has narrower G-band line
width and a lower D/G ratio indicating the higher purity of this
sample in comparison with the parent suspension.
PARENT
UCF
ATP Bottom
ATP Top
(9,1)
(9,1)*
(10,2)*
ex
= 725nm
(9,1)*
210 240 270 300
Raman Shift (cm-1)
ex
= 676.4nm, HiPco 195.2
Raman Shift (cm-1)
Norm
aliz
ed
Ram
an
Inte
nsity (
a.u
.)
(b)
(10,2)
ex
= 676.4nm
ex
= 785nm
(a)
G
PARENT (0.02)
UCF (0.015)
DGU (0.015)
ATP Bottom (0.015)
ATP Top (0.027)
PARENT (0.07)
UCF (0.036)
DGU (0.034)
ATP Bottom (0.034)
ATP Top (0.08)
ex
= 514.5nm, HiPco 195.2
1300 1400 1500 1600
D
PARENT (0.018)
ATP Bottom (0.01)
ATP Top (0.036)
ex
= 514.5nm, Plasma Torch
| www.editorialmanager.com/nare/default.asp
6 Nano Res.
greater absorption of the UCF dispersion, relative to
ATP and DGU, corroborates the presence of more
impurities and bundles contributing to the overall
absorption. Raman spectroscopy, Atomic Force
Microscopy (AFM), Transmission Electron
Microscopy (TEM), and Thermogravimetric Analysis
(TGA) were used to further evaluate the relative
purity of the ATP extracted SWCNTs. SWCNT
isolation is confirmed by Raman spectra (Fig. 2a)
obtained at 785 nm excitation, where bundling shifts
the resonance conditions for the (10,2)-SWCNT (d=
0.884 nm) from 735 nm to 785 nm, making its radial
breathing mode at 266cm-1 more prominent as
bundle content increases [12]. Given that
(10,2)-SWCNTs might not be present in all source
materials, we searched for a new SWCNT family to
probe bundle formation. The (9,1) SWCNT (d= 0.757
nm) normally resonant at 680 nm becomes resonant
at 725 nm when bundled and serves as an alternative
probe for bundle identification. Thus, the presence of
the Raman bundling peaks reveals that unlike either
ATP or DGU, the absorption spectra of the
dispersions obtained by CF and UCF maintain a
significant background due to the presence of
bundles (Fig. 2). The Raman D/G ratios at different
excitation wavelengths (Fig. 2b) shows that the
G-band line-width and D/G ratios reduce similarly
for ATP and DGU (relative to the parent and UCF
samples), indicating highly isolated samples after
ATP separation [26, 27].
AFM (Fig. 3a) corroborates minimal bundles in the
bottom phase, and indicates that most bundles
remain in the top and interfacial layers. Statistical
analysis of the AFM data (Fig. 3b) reveals at least 90%
of the SWCNT present in the bottom phase are well
isolated or not bundled, some of which were created
during AFM sample preparation. TGA confirms a
decrease in catalyst concentration after ATP
extraction, corroborated by TEM (see Fig. S2), from
26.4% in the starting raw sample to 12.6% in the
bottom phase (similar to UCF and magnetic
purifications [14], Fig. S3). Since ATP and DGU are
equilibrium-based separations, unlike UCF or
magnetic purification that are rate-based, a greater
isolation from impurities can be achieved without
sacrificing yield.
Figure 3: AFM analysis of the separated fractions. (a)
Representative AFM image of the top and bottom phase after
bundle extraction for the parent HiPco 195.2 dispersion, showing
many more impurities and bundles in the top phase and (b) height
statistics of the AFM images, revealing 90% of isolated SWCNTs
in the bottom phase (note that HiPco diameters range from
~0.54-1.2nm[25] and that AFM sample deposition might cause a
small degree of bundling). The inset zooms in on the relevant
region for the bottom phase.
The mechanism of ATP separation is directly dictated
by its constituent hydrophobicities [23], not intrinsic
SWCNT hydrophobicity [8]. As such, parameters that
are not directly related to the hydrophobicity of the
SWCNT/micelle system, all yield similar SWCNT
extractions (Fig. 4a-b and Fig. S4). For example, no
change in isolation quality is apparent when
increasing the sonication time (Fig. 4a) or the
SWCNT starting concentration (Fig. S4 in ESM),
which only result in a higher overall concentration of
isolated SWCNTs. As well, an additional
centrifugation step prior to the ATP separation to
partially remove bundles only slightly reduces the
separation yield (Fig. 4b). For parameters that affect
the relative hydrophobicity of the two phases
however, a different SWCNT isolation is achieved
(Fig. 4c-d, Fig. S5 and S6 in ESM). However, the
versatility of the method allows one to compensate
these small changes with small changes of the DOC
concentration used for the separation; decreasing
DOC concentration until optimal, increases the
number of isolated SWCNTs in the bottom phase,
while increasing DOC concentrations until optimal,
0 5 10 15 20
0
5
10
15
20
ATP Bottom
Co
unt
ATP Top
0 5 10 15 20
0
20
40
60
Co
unt
Height (nm)
0 1 2 3 4
0
20
40
(b)
(a)
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
7 Nano Res.
Figure 4: Parameters that affect and those that do not affect the
separation. Not Affecting separation: (a) Variation of sonication
time: As measured absorption spectra of the ATP bottom phases
for a HiPco 195.2 parent suspension prepared with either 1 or 24
hours of sonication. After rescaling (inset), no significant change
in the chirality distribution nor in the ratio of background to
SWCNT signals is found. (b) As measured absorption spectra of
the ATP bottom phase for two different parent suspensions, with
and without a pre-centrifugation step. Only a slight overall
decrease of the signals, no change in ratio background versus
SWCNT signals (see inset for rescaled absorption spectra).
Affecting Separation: (c) Volumetric diluted absorption spectra of
the ATP-separated bottom phases for different dextran MW, with
the inset presenting the rescaled absorption spectra. (d) As
measured absorption spectra of the ATP separated bottom phase
for different separation temperatures. At high temperature,
individualized SWCNTs move to the top phase, while at low
temperature background increases due to more bundles and
impurities separating in the bottom phase. All separations were
performed using a fixed DOC concentration of 0.088%wt/V.
removes bundles from the bottom phase (Fig. 1b).
intrinsic SWCNT hydrophobicity [8]. As such,
parameters that are not directly related to the
hydrophobicity of the SWCNT/micelle system, all
yield similar SWCNT extractions (Fig. 4a-b and Fig.
S4). For example, no change in isolation quality is
apparent when increasing the sonication time (Fig. 4a)
or the SWCNT starting concentration (Fig. S4 in
ESM), which only result in a higher overall
concentration of isolated SWCNTs. As well, an
additional centrifugation step prior to the ATP
separation to partially remove bundles only slightly
reduces the separation yield (Fig. 4b). For parameters
that affect the relative hydrophobicity of the two
phases however, a different SWCNT isolation is
achieved (Fig. 4c-d, Fig. S5 and S6 in ESM). However,
the versatility of the method allows one to
compensate these small changes with small changes
of the DOC concentration used for the separation;
decreasing DOC concentration until optimal,
increases the number of isolated SWCNTs in the
bottom phase, while increasing DOC concentrations
until optimal, removes bundles from the bottom
phase (Fig. 1b).
Albertsson et al. [23] showed that by lowering the
molecular weight of dextran, the phase separation
and thus the redistribution of the particles across the
two phases is altered with more particles separating
into the bottom phase. Likewise, in our separation,
lower dextran MW results in a small fraction of
bundles reaching the bottom phase (mol. wt. 9000Da),
which can be corrected by decreasing the DOC
concentration (Fig. 4c and Fig. S5 in ESM), resulting
in an identical quality and distribution of isolated
SWCNTs. Such control permits the use of lower
molecular weight and low cost industrial grade
dextran, which would be beneficial for ease of
dialysis after separation (to remove the polymers
from the separated fractions) and to reduce the
separation cots. Similar to other separation
techniques [28], ATP separations are sensitive to
temperature [19], due to its significant effect on the
specific surfactant structure [29], thereby altering the
hydrophobicity. Typically, lower temperature results
in more bundles in the bottom phase, while higher
temperature results in less isolated SWCNTs in the
bottom phase, but both can be corrected by altering
the DOC concentration (Fig. 4d).
We furthermore prepared SWCNT dispersions and
performed a similar separation using sodium
dodecylsulfate (SDS) and sodium cholate (SC) as the
dispersing agent in the same PEG/dextran stock
solution (Fig. S6). SDS is unable to extract only
individualized SWCNTs, with bundles and most
isolated SWCNTs separating into the same (top)
phase; the small amount that does remain in the
bottom phase contains bundles. With SC, individual
SWCNTs can be isolated; however, a higher
300 600 900 1200 15000.0
0.2
0.4
0.6
0.8
300 600 900 1200
300 600 900 1200
300 600 900 1200 15000.0
0.2
0.4
0.6
0.8
1.0
1.2
300 600 900 1200
300 600 900 1200 15000.0
0.1
0.2
0.3
300 600 900 1200
(a) (b)
(c) (d)
Ab
s (
1m
m)
Sonication Time
24h
1h
Wavelength (nm)
300 600 900 1200 15000.0
0.1
0.2
0.3
Ab
s (
1m
m)
Additional CF step
no CF
CF
Ab
s (
1m
m)
Temperature
7°C
22°C
50°C
Wavelength (nm)
Wavelength (nm)
Ab
s (
1m
m)
Dextran MW
~70 KDa
~35 KDa
~9 KDa
Wavelength (nm)
Not Affecting Separation Affecting Separation
| www.editorialmanager.com/nare/default.asp
8 Nano Res.
concentration of surfactant is needed than with DOC,
since at lower concentrations, the resulting
distribution of isolated SWCNTs is skewed toward
small-diameter SWCNTs. Briefly, the differences in
surfactant behaviors are due to the dynamic nature
of their wrapping configurations, which can be
linked to their molecular structure as previously
discussed in detail [20].
For a particular separation technique to be of high
impact, it is crucial to demonstrate that it is widely
applicable for a variety of SWCNT source materials.
Different SWCNT production methods (Table S1 in
ESM), including less expensive sources, offer variable
diameter distributions, production related defect
densities, and SWCNT content (e.g. Arc-discharge
synthesis typically yields <40% SWCNT content
[27]). Importantly, the dominant physical parameters
of the SWCNTs (defect density, diameter, and length)
should not affect the constituent hydrophobicities
and thus the quality of isolation.
Figure 5: ATP isolation of SWCNTs from different sources
(parent suspensions prepared in D2O for optical transparency in
the NIR). Absorption spectra of the parent (black) and resulting
ATP bottom fractions (red) for (a) Laser vaporization SWCNTs;
(b) Plasma Torch (Raymor) SWCNTs; (c) Arc-discharge
SWCNTs and (d) CoMoCat SG65 SWCNTs. Absorption spectra
are presented for 1mm path length, and ATP bottom fractions are
rescaled according to the volumetric dilution. The insets of each
panel present a zoomed-in absorption of the ATP-sorted bottom
fractions in comparison with a rescaled absorption of the parent
suspension (long-wavelengths normalization). The scaling factors
for the parent suspension are also given. Data are cut around
2000nm, due to small H2O absorption in the D2O-sorted samples.
Figure 5 shows the absorption spectra of samples,
before and after ATP separation, of various SWCNT
source materials used across the vast SWCNT
literature.
ATP successfully removes impurities from all sources
exhibiting the same background reduction shown in
HiPco dispersions. To better assess the changes in
spectra, we scaled the absorption of the parent
dispersions such that both spectra coincide at high
wavelengths (i.e. rescaling on the non-resonant
carbon absorbance, insets of Fig. 5a-d). Spectral
resolution for all ATP isolated SWCNT samples
exceed that of the parent suspensions, revealing the
excitonic transitions that are overshadowed in the
parent samples by strong backgrounds attributable to
the presence of impurities and bundles. AFM images
of the top and bottom phase of Plasma Torch CNTs
(Fig. S7 in ESM), TEM (Fig. S8 in ESM) and RRS
spectra (Fig. 2b) confirm mostly isolated SWCNTs are
extracted to the bottom phase while carbonaceous
impurities and bundles separate into the top phase.
To better evaluate the quality of SWCNT isolation
from a low cost source, we compared spectra from
parent, centrifuged, and ATP dispersions of Plasma
Torch SWCNTs from Raymor Technologies. A
relative PL efficiency is presented in Fig. 6a, where
PL measurements were scaled by respective sample
absorbance at the 800nm excitation wavelength.
Upon centrifugation, SWCNT fluorescence increases
by a factor of two, whereas ATP separated
dispersions achieve an order of magnitude greater PL
for a given sample absorption (absorption at 800nm).
The quality of the isolated Plasma Torch SWCNTs is
highlighted by the PLE map identifying emitting
SWCNTs with diameters from ~1-1.5nm (Fig. 6b).
Figure 6: (a) Fluorescence spectra of the Plasma Torch SWCNTs
from Raymor parent (black), centrifuged (blue) and ATP sorted
bottom phase (red). PL intensity is scaled by the absorbance at
500 1000 1500 20000
2
4
6
8
500 1000 1500 20000
5
10
15
500 1000 1500 20000.0
0.2
0.4
0.6
0.8
500 1000 1500 20000
1
2
3
500 1000 1500 20000
1
2
3
500 1000 1500 20000
1
2
3
Ab
s (
1m
m)
Plasma Torch PARENT
ATP
Wavelength (nm)
500 1000 1500 20000
2
4
6
8
Ab
s (
1m
m)
Arc-Discharge PARENT
ATP
500 1000 1500 20000
1
2
3
4
5
6
7
8
Wavelength (nm)
(d)
(b)
(c) CoMoCat SG65 PARENT
ATP
Ab
s (
1m
m)
Wavelength (nm)
(a)
Laser Vaporization PARENT
ATP
Ab
s (
1m
m)
Wavelength (nm)
PARENT / 4 PARENT / 2.7
PARENT / 5.5
PARENT / 12
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
9 Nano Res.
the excitation wavelength (800nm) to obtain a relative dispersion
PL efficiency. (b) 2D NIR PLE map of Plasma Torch ATP
separated bottom fraction, showing excitonic transitions of
SWCNTs in the 1-1.5nm diameter range. The PLE spectrum was
calibrated for excitation light intensity and detector sensitivity.
Likewise, the bottom fractions in all tested source
materials exhibit well-resolved PL peaks indicative of
high-quality SWCNT isolation (Fig. S9). Recent work
has furthermore shown that the PL can even be
further enhanced with plasmonic colloidal
suspensions [30]. Thus, separation of SWCNTs
from various sources offers a large range of isolated
diameters, which are instrumental for enabling a
broad range of SWCNT applications. For instance,
while SWCNT photovoltaics show better
performance with small diameter SWCNTs [31],
transistors require large diameters [32, 33], and
biological imaging needs diameters (i.e. band-gap)
tuned [2, 34, 35] to avoid absorption by water, tissue,
etc..
Besides the applicability of the method for a wide
range of source materials, the scalability and
repeatability of the method is extremely important
for a new purification method to meet application
needs. We therefore performed separations on a 0.472
mL, a 4.72 mL and a 47.2 mL scale, thus generating
bottom fractions from 90 µL to 9 mL in the same time
frame. All separations achieve identical separations
irrespective of 2 orders of magnitude change in
volume scale (Figure 7).
Figure 7: Scalability of the ATP separation: (a) As-measured
absorption spectra of the ATP bottom fractions after separating in
3 different volumes, resulting in 90µL, 900µL and 9mL of
bottom fraction. (b) Photographs of the resulting
phase-separations.
Finally, large-scale isolation of pristine, full-length
SWCNTs has remained elusive in the SWCNT
community, as most separation methods demand
intense sonication [2, 3, 5], which reduces the original
aspect ratio (full-length). Various SWCNT
applications, such as conductive/strong fibers [36],
thin-film transistors [32], strain sensors [37] and
conductive transparent films [38, 39], as well as
experimental studies [40] have stressed the
importance of long, pristine SWCNTs to enhance
mechanical strength, electrical/thermal conductivity
Figure 8: ATP separation of stirred SWCNT samples (no
sonication; long SWCNTs). (a) Absorption spectra of the parent
(black) and ATP bundle-extracted bottom phase (as according to
its volumetric dilution) (red) for HiPco 195.2 SWCNT dispersion,
prepared without sonication. The inset presents an expanded
absorption spectrum of the ATP bottom phase. (b) RRS spectra
excited at 570nm for the starting HiPco 195.2 stirred parent
solution, and the ATP sorted bottom phase. Red and blue dashed
lines indicate positions of empty and water-filled (H2O@CNT)
SWCNTs, respectively. Ratios between empty and water-filled
RBM intensities remain constant after separation. (c-d)
representative set of PL images (excitation 570 nm, CCD
detection) of the parent (c) and ATP bottom (d) phase after
bundle-extraction, clearly showing both short and long SWCNTs
are separated in a similar way.
and optical quality. Full-length SWCNTs not only
have higher aspect ratios, but are also less defective
and display improved optical properties [41, 42].
Only a few studies have reported isolation of long
SWCNTs by DGU [42, 43], but at the cost of
simplicity, time and scalability. To generate long,
pristine SWCNT dispersions using our
ATP-technique, the SWCNT-surfactant mixture is
gently stirred over a period of 1-3 weeks [4], typically
300 600 900 1200 15000.0
0.2
0.4
300 600 900 1200 15000.00
0.02
300 310 320 330 340
(b)
(d)
HiPco 195.2 (no sonication)
PARENT
ATP
Ab
s (
1m
m)
Wavelength (nm)
(a)
(c)
(6,5)(6,4)
H2O@(6,4)
ATP Bottom
Ra
ma
n I
nte
nsity (
a.u
.)
Raman shift (cm-1)
Parent
H2O@(6,5)
| www.editorialmanager.com/nare/default.asp
10 Nano Res.
resulting in a mixture of closed ('pristine') and
opened ('cut') SWCNTs with few defects [7, 25, 41,
42]. Figure 8a (and Fig. S10 in ESM) demonstrates
that ATP can be used to successfully extract isolated
SWCNTs from these dispersions (red trace), resulting
in a significant reduction of the background
absorption and sharp excitonic transitions, which are
not resolved in the parent dispersion (black trace).
The significant changes in the background and
excitonic linewidth are directly associated with the
lower SWCNT isolation obtained by simple stirring,
however by stirring a higher SWCNT concentration,
very similar optical density of the resulting bottom
fractions can be obtained (comparing insets of Fig. 1b
with Fig. 8a). PL images (Figure 8c-d) of the isolated
SWCNTs and parent dispersion show that SWCNTs
up to 20 µm in length remain after ATP separation.
AFM measurements (Fig. S11) confirm that mostly
isolated long SWCNTs are present in the bottom
phase, while the majority of bundles and other
impurities remain in the top phase. A further
measure of the gentleness of this extraction technique
can also be demonstrated by high resolution RRS,
which allows resolving the presence of closed (empty,
pristine) and opened (water-filled) SWCNTs, since
the water-filled SWCNTs have slightly blue-shifted
RBM vibrations [7, 25, 41, 42]. Figure 8b presents
such Raman spectra excited at 570nm for the stirred
HiPco parent solution and the ATP sorted bottom
phase, clearly indicating that a significant fraction of
the SWCNTs are closed since no sonication was
applied, and that the ratio of closed-to-opened
SWCNTs remains the same after separation. These
results demonstrate that ATP multi-chiral separation
can be achieved, independent of the length and
average diameter of SWCNTs.
Similarly, we have applied our previous
single-chirality (6,5) separation approach [20] to these
stirred SWCNT samples, while also avoiding a prior
UCF step, to achieve a fast single-chirality separation
of full-length SWCNTs. A similar distribution of
SWCNT chiralities is achieved with minimal
background in absorption spectra for both sonicated
(1h) and stirred samples (Figure 9a). Notably,
sonication increases the overall concentration of
isolated SWCNTs in the parent solution and thus a
greater amount of (6,5)-SWCNT enrichment after
separation, though at the cost of shortening. Figures
9b and 9c show single-SWCNT PL images after ATP
(6,5) enrichment of long and short SWCNTs,
respectively, revealing that even one hour of
sonication generates samples with SWCNTs no
longer than 1µm while stirring maintains full-length
SWCNTs (up to 20µm in the studied samples).
Moreover, this confirms that the chirality separation
is determined by the specific surfactant composition
covering the SWCNTs, independent of their length.
Figure 9: Single-chirality (6,5) ATP separation [20] for short and long SWCNTs. (a) As-measured absorption spectra of (6,5) sorted
bottom phases for stirred (no sonication, black) and tip-sonicated (1 hour, red) HiPco 195.2 SWCNT dispersion after ATP isolation.
(b-Cc) Representative set of PL images (excitation 570 nm, CCD detection) of the sonicated (b) and non-sonicated (c) (6,5) fractions
showing the influence of sonication on the length distribution in the samples.
Availability of long pristine SWCNTs should find
immediate applications in conductive transparent
thin-films and aligned arrays of long SWCNTs for
technologies such as photovoltaics, video displays,
transistors, conductive/strong fibers, and solid-state
lighting. The performance of such devices should
benefit tremendously from very long SWCNTs, since
longer SWCNTs lower the percolation threshold and
can enhance SWCNT alignment over long distances,
which in turn lowers the resistance of thin-films,
increases the strength and conductivity of fibers, and
increases carrier mobility, on-currents, and on/off
300 600 900 12000.00
0.05
0.10
no sonication
sonication
Ab
s (
1m
m)
Wavelength (nm)
(a) (b) (c)
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
11 Nano Res.
ratios of transistors [39, 44].
4. Conclusions.
Here we report multi-chirality isolation of individual
SWCNTs through a simple ATP technique. Our
technique achieves DGU quality dispersions in a
timely, inexpensive, and highly accessible manner
attaining low background absorption and sharp
excitonic transitions, due to the removal of starting
impurities and bundles. With such purified samples
we were able to identify a new resonance condition
(725 nm) where the (9,1) SWCNT-family comes into
resonance as a result of bundling serving as an
alternative standard (to the typically utilized (10,2)
standard) to probe bundle presence.
Remarkably, this technique extracts high-quality
isolated SWCNT dispersions from any SWCNT
source, irrespective of purity, diameters (tested up to
1.6 nm), opened/closed ratio, and length (tested up to
20 μm), substantiating our claim that the nature of
the nanotubes does not directly affect the
hydrophobicity of the system and thus the quality of
isolation. Moreover, controlling the constituents
affecting hydrophobicity enables tunability toward
desired multi- and single-chirality isolations. Such
control enables the use of cheap SWCNTs (e.g.
Plasma Torch (Raymor)) to achieve inexpensive
high-quality isolated SWCNT dispersion. Future
studies will explore variation of our techniques
toward isolation of larger diameter SWCNT,
few-walled carbon nanotubes, open/closed
separation, and metal/semiconductor separation and
development of a procedure to accurately measure
the overall yield of all separations.
We anticipate that as observed in the surge of
graphene research that followed the introduction of
solution phase processing [45], our ability to create
high-quality suspensions via high-throughput,
inexpensive, and gentle methods from low-cost
materials without expensive instruments (replacing
UCF by bench-top centrifugation and potentially
without any sonication) will kindle new SWCNT
research in multidisciplinary fields, promoting
innovation in fundamental research and
high-performance devices. Furthermore, our
approach will accelerate the incorporation of isolated
long SWCNTs from low-cost sources into rapidly
emerging commercial applications such as material
reinforcement, sensors, field-effect transistors, and
photovoltaics where large quantities of full-length
CNTs are required in a cost effective manner.
Acknowledgements
We thank Kevin Henderson for access to TGA and
Wim Wenseleers for helpful discussions. This work
was supported by the LANL-LDRD program and
was performed in part at the Center for Integrated
Nanotechnologies, a U.S. Department of Energy,
Office of Basic Energy Sciences user facility. S.C.
gratefully acknowledges the financial support from
the Fund for Scientific research Flanders, Belgium
(FWO-Vlaanderen) for providing a postdoctoral
fellowship and a mobility grant for visiting the Los
Alamos National Laboratory. A.N.G.P.V. gratefully
acknowledges support from the LANL Director’s
Postdoctoral Fellowship. LV SWCNT synthesis was
performed at NREL, and was supported by the Solar
Photochemistry Program, Division of Chemical
Sciences, Geosciences, and Biosciences, Office of
Basic Energy Sciences, U.S. Department of Energy
(DOE), Grant DE-AC36-08GO28308..
Electronic Supplementary Material: Supplementary
material (description of materials, methods used for
the spectroscopic and image characterization, and
supporting figures as referred to in the manuscript)
is available in the online version of this article at
http://dx.doi.org/10.1007/s12274-***-****-*. References
[1] Jorio, A.; Dresselhaus, G.; Dresselhaus, M. S.
Carbon Nanotubes: Advanced Topics in
Synthesis, Structure, Properties and
Applications; Springer-Verlag: Berlin, 2008.
[2] O'Connell, M. J.; Bachilo, S. M.; Huffman, C.
B.; Moore, V. C.; Strano, M. S.; Hároz, E. H.;
Rialon, K. L.; Boul, P. J.; Noon, W. H.;
Kittrell, C.; et al. Band gap fluorescence
from individual single-walled carbon
nanotubes. Science 2002, 297, 593-596.
[3] Islam, M. F.; Rojas, E.; Bergey, D. M.;
| www.editorialmanager.com/nare/default.asp
12 Nano Res.
Johnson, A. T.; Yodh, A. G. High weight
fraction surfactant solubilization of
single-wall carbon nanotubes in water. Nano
Lett. 2003, 3, 269-273.
[4] Wenseleers, W.; Vlasov, I.I.; Goovaerts, E.;
Obraztsova, E. D.; Lobach, A. S.; Bouwen, A.
Efficient isolation and solubilization of
pristine single-walled nanotubes in bile salt
micelles. Adv. Funct. Mater. 2004, 14,
1105-1112.
[5] Cathcart, H.; Quinn, S.; Nicolosi, V.; Kelly, J.
M.; Blau, W. J.; Coleman, J. N. Spontaneous
debundling of single-walled carbon
nanotubes in DNA-Based dispersions. J.
Phys. Chem. C 2006, 111, 66-74.
[6] Gerstel, P.; Klumpp, S.; Hennrich, F.;
Altintas, O.; Eaton, T. R.; Mayor, M.;
Barner-Kowollik, C.; Kappes, M. M.
Selective dispersion of single-walled carbon
nanotubes via easily accessible conjugated
click polymers. Polym. Chem. 2012, 3,
1966-1970.
[7] Wenseleers, W.; Cambré, S.; Culin, J.;
Bouwen, A.; Goovaerts, E. Effect of water
filling on the electronic and vibrational
resonances of carbon nanotubes:
characterizing tube opening by Raman
spectroscopy. Adv. Mater. 2007, 19,
2274-2278.
[8] Parra-Vasquez, A. N. G.; Duque, J. G.;
Green, M. J.; Pasquali, M. Assessment of
length and bundle distribution of dilute
single-walled carbon nanotubes by
viscosity measurements. AIChE Journal 2014,
60, 1499-1508.
[9] Bonaccorso, F.; Hasan, T.; Tan, P. H.; Sciascia,
C.; Privitera, G.; Di Marco, G.; Gucciardi, P.
G.; Ferrari, A. C. Density gradient
ultracentrifugation of nanotubes: Interplay
of bundling and surfactants encapsulation. J.
Phys. Chem. C 2010, 114, 17267-17285.
[10] Arnold, M. S.; Stupp, S. I.; Hersam, M. C.
Enrichment of single-walled carbon
nanotubes by diameter in density gradients.
Nano Lett. 2005, 5, 713-718.
[11] Tabakman, S. M.; Welsher, K.; Hong, G.; Dai,
H. Optical properties of single-walled
carbon nanotubes separated in a density
gradient: Length, bundling, and aromatic
stacking effects. J. Phys. Chem. C 2010, 114,
19569-19575.
[12] O’Connell, M. J.; Sivaram, S.; Doorn, S. K.
Near-Infrared resonance Raman excitation
profile studies of single-walled carbon
nanotube intertube interactions: A direct
comparison of bundled and individually
dispersed HiPco nanotubes. Phys. Rev. B
2004, 69, 235415.
[13] Arnold, M. S.; Green, A. A.; Hulvat, J. F.;
Stupp, S. I.; Hersam, M. C. Sorting carbon
nanotubes by electronic structure using
density differentiation. Nature Nanotechnol.
2006, 1, 60-65.
[14] Ghosh, S.; Bachilo, S. M.; Weisman, R. B.
Removing aggregates from single-walled
carbon nanotube samples by magnetic
purification. J. Phys. Chem. C 2014, 118,
4489-4494.
[15] Arnold, M. S.; Suntivich, J.; Stupp, S. I.;
Hersam, M. C. Hydrodynamic
characterization of surfactant encapsulated
carbon nanotubes using an analytical
ultracentrifuge. ACS Nano 2008, 2,
2291-2300.
[16] Naumov, A. V.; Kuznetsov, O. A.;
Harutyunyan, A. R.; Green, A. A.; Hersam,
M. C.; Resasco, D. E.; Nikolaev, P. N.;
Weisman, R. B. Quantifying the
semiconducting fraction in single-walled
carbon nanotube samples through
comparative atomic force and
photoluminescence microscopies. Nano Lett.
2009, 9, 3203-3208.
[17] Fagan, J. A.; Zheng, M.; Rastogi, V.;
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
13 Nano Res.
Simpson, J. R.; Khripin, C. Y.; Batista, C. A.
S.; Walker, A. R. H. Analyzing surfactant
structures on length and chirality resolved
(6,5) single-wall carbon nanotubes by
Analytical Ultracentrifugation. ACS Nano
2013, 7, 3373-3387.
[18] Khripin, C. Y.; Fagan, J. A.; Zheng, M.
Spontaneous partition of carbon nanotubes
in polymer-modified aqueous phases. J. Am.
Chem. Soc. 2013, 135, 6822-6825.
[19] Fagan, J. A.; Khripin, C. Y.; Silvera Batista, C.
A.; Simpson, J. R.; Hároz, E. H.; Hight
Walker, A. R.; Zheng, M. Isolation of specfic
small diameter single-wall carbon nanotube
species via aqueous two-phase extraction.
Adv. Mater. 2014, 26, 2800-2804.
[20] Subbaiyan, N. K.; Cambré, S.;
Parra-Vasquez, A. N. G.; Hàroz, E. H.;
Doorn, S. K.; Duque, J. G. Role of
surfactants and salt in aqueous two-phase
separation of carbon nanotubes toward
simple chirality isolation. ACS Nano 2014, 8,
1619-1628.
[21] Zhang, M.; Khripin, C. Y.; Fagan, J. A.;
McPhie, P.; Ito, Y.; Zheng, M. Single-step
total fractionation of single-wall carbon
nanotubes by countercurrent
chromatography. Anal. Chem. 2014, 86,
3980-3984.
[22] Ao, G.; Khripin, C. Y.; Zheng, M.
DNA-controlled partition of carbon
nanotubes in polymer aqueous two-phase
systems. J. Am. Chem. Soc. 2014, 136,
10383-10392.
[23] Albertsson, P.-A. Partition of Cell Particles
and Macromolecules: Separation and
Purification of Biomolecules, Cell Organelles,
Membranes and Cells in Aqueous Polymer Two
Phase Systems and Their Use in Biochemical
Analysis and Biotechnology.; John Wiley and
Sons, 1986.
[24] Duque, J. G.; Pasquali, M.; Cognet, L.;
Lounis, B. Environmental and
synthesis-dependent luminescence
properties of individual single-walled
carbon nanotubes. ACS Nano 2009, 3,
2153-2156.
[25] Cambré, S.; Schoeters, B.; Luyckx, S.;
Goovaerts, E.; Wenseleers, W. Experimental
observation of single-file water filling of
thin single-wall carbon nanotubes down to
chiral index (5,3). Phys. Rev. Lett. 2010, 104,
207401.
[26] Dresselhaus, M. S.; Jorio, A.; Souza Filho, A.
G.; Saito, R. Defect characterization in
graphene and carbon nanotubes using
Raman spectroscopy. Philos. Trans. R. Soc., A
2010, 368, 5355-5377.
[27] Miyata, Y.; Mizuno, K.; Kataura, H. Purity
and defect characterization of single-wall
carbon nanotubes using Raman
spectroscopy. J. of Nanomater. 2011, 2011.
[28] Liu, H.; Tanaka, T.; Urabe, Y.; Kataura, H.
High-efficiency single-chirality separation
of carbon nanotubes using
temperature-controlled gel chromatography.
Nano Lett. 2013, 13, 1996-2003.
[29] D'Alagni, M.; D'Archivio, A. A.; Galantini,
L.; Giglio, E. Structural study of the micellar
aggregates of sodium chenodeoxycholate
and sodium deoxycholate. Langmuir 1997,
13, 5811-5815.
[30] Glaeske, M.; Setaro, A. Nanoplasmonic
colloidal suspensions for the enhancement
of the luminescent emission from
single-walled carbon nanotubes. Nano Res.
2013, 6, 593-601.
[31] Shea, M. J.; Arnold, M. S. 1% solar cells
derived from ultrathin carbon nanotube
photoabsorbing films. Appl. Phys. Lett. 2013,
102, 243101.
[32] Wang, C.; Zhang, J.; Ryu, K.; Badmaev, A.;
De Arco, L. G.; Zhou, C. Wafer-scale
fabrication of separated carbon nanotube
| www.editorialmanager.com/nare/default.asp
14 Nano Res.
thin-film transistors for display applications.
Nano Lett. 2009, 9, 4285-4291.
[33] Zhang, J.; Gui, H.; Liu, B.; Liu, J.; Zhou, C.
Comparative study of gel-based separated
arcdischarge, HiPCO, and CoMoCAT
carbon nanotubes for macroelectronic
applications. Nano Res. 2013, 6, 906-920.
[34] Diao, S.; Hong, G.; Robinson, J. T.; Jiao, L.;
Antaris, A. L.; Wu, J. Z.; Choi, C. L.; Dai, H.
Chirality enriched (12,1) and (11,3)
single-walled carbon nanotubes for
biological imaging. J. Am. Chem. Soc. 2012,
134, 16971-16974.
[35] Hong, G.; Lee, J. C.; Robinson, J. T.; Raaz, U.;
Xie, L.; Huang, N. F.; Cooke, J. P.; Dai, H.
Multifunctional in vivo vascular imaging
using near-infrared Fluorescence. Nature
Med 2012, 18, 1841-1846.
[36] Behabtu, N.; Young, C. C.; Tsentalovich, D.
E.; Kleinerman, O.; Wang, X.; Ma, A. W. K.;
Bengio, E. A.; ter Waarbeek, R. F.; de Jong, J.
J.; Hoogerwerf, R. E.; et al. Strong, light,
multifunctional fibers of carbon nanotubes
with ultrahigh conductivity. Science 2013,
339, 182-186.
[37] Wu, M.; Liu, K.; Wang, W.; Sui, Y.; Bai, X.;
Wang, E. Ultralong aligned single-walled
carbon nanotubes on flexible
fluorphlogopite mica for strain sensors.
Nano Res. 2012, 5, 443-449.
[38] Tenent, R. C.; Barnes, T. M.; Bergeson, J. D.;
Ferguson, A. J.; To, B.; Gedvilas, L. M.;
Heben, M. J.; Blackburn, J. L. Ultrasmooth,
large-area, high-uniformity, conductive
transparent single-walled carbon nanotube
films for photovoltaics produced by
ultrasonic spraying. Adv. Mater. 2009, 21,
3210-3216.
[39] Wang, Y.; Huang, L.; Liu, Y.; Wei, D.; Zhang,
H.; Kajiura, H.; Li, Y. Minimizing
purification-induced defects in
single-walled carbon nanotubes gives films
with improved conductivity. Nano Res. 2009,
2, 865-871.
[40] Zhao, S.; Kitagawa, T.; Miyauchi, Y.;
Matsuda, K.; Shinohara, H.; Kitaura, R.
Rayleigh scattering studies on inter-layer
interactions in structure-defined individual
double-wall carbon nanotubes. Nano Res.
2014, 7, 1548-1555.
[41] Cambré, S.; Santos, S. M.; Wenseleers, W.;
Nugraha, A. R. T.; Saito, R.; Cognet, L.;
Lounis, B. Luminescence properties of
individual empty and water-filled
single-walled carbon nanotubes. ACS Nano
2012, 6, 2649-2655.
[42] Cambré, S.; Wenseleers, W. Separation and
diameter-worting of empty (end-capped)
and water-filled (open) carbon nanotubes
by density gradient ultracentrifugation.
Angew. Chem. - Int. Ed. 2011, 50, 2764-2768.
[43] Fagan, J. A.; Becker, M. L.; Chun, J.; Hobbie,
E. K. Length fractionation of carbon
nanotubes using centrifugation. Adv. Mater.
2008, 20, 1609-1613.
[44] Wu, Z.; Chen, Z.; Du, X.; Logan, J. M.;
Sippel, J.; Nikolou, M.; Kamaras, K.;
Reynolds, J. R.; Tanner, D. B.; Hebard, A. F.;
et al. Transparent, conductive carbon
nanotube films. Science 2004, 305, 1273-1276.
[45] Zhu, Y.; Murali, S.; Cai, W.; Li, X.; Suk, J. W.;
Potts, J. R.; Ruoff, R. S. Graphene and
graphene oxide: synthesis, properties, and
applications. Adv. Mater. 2010, 22,
3906-3924.
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
Nano Res.
Electronic Supplementary Material
Benchtop aqueous two-phase extraction of isolated
individual single-walled carbon nanotubes
Navaneetha K. Subbaiyan,1,† A. Nicholas G. Parra-Vasquez, 1,† Sofie Cambré,1,2,† Miguel A. Santiago
Cordoba,1 Sibel Ebru Yalcin,1 Christopher E. Hamilton,1 Nathan H. Mack,1 Jeffrey L. Blackburn,3 Stephen K.
Doorn,1 and Juan G. Duque1().
Supporting information to DOI 10.1007/s12274-****-****-* (automatically inserted by the publisher)
S1: Materials
Table – S1: – Various SWCNT source materials used in this study. The diameter range was extracted from
absorption spectra.
Name Manufacturer Batch Diameter range
HiPco Rice University 195.2 d=0.9±0.3 nm
CoMoCat SG65i SWeNT SG65i-L39 d=0.7±0.1 nm
Laser Vaporization NREL Y140307 d=1.2±0.2 nm
Plasma Torch Raymor RNL12-010-113 d=1.3±0.2 nm
Arc-discharge 1 Simga Aldrich 07826BA d=1.4±0.2 nm
Arc-discharge 2 Nanoledge P00508D d=1.5±0.2 nm
S2: Methods
S2.a. Optical Absorption Spectroscopy and Volume Dilution Correction
Absorption spectra were recorded with a UV-vis-NIR spectrophotometer (Cary 6000i) in the range of
200-1700nm (H2O-samples) and with a Cary 5E spectrophotometer in the range of 175-2500nm (D2O-samples),
using fused silica cells with an optical path length of 1 cm or 3 mm. Spectra were diluted if necessary (to keep
OD < 3) and afterwards rescaled according to their dilution factor and to a 1mm optical path length.
For proper comparison of the separation yields, absorption spectra were rescaled according to their
volumetric dilutions as in our previous work.[20]
We start with a mass balance:
| www.editorialmanager.com/nare/default.asp
Nano Res.
(1)
where is the mass of the SWCNTs in the parent dispersion (P), top phase (T), bottom phase (B) and the
interface (I). The mass can be determined from absorption through an extinction coefficient, ε:
(2)
where is absorption and is the volume of the respective dispersion. Assuming the extinction
coefficient does not change significantly in each fraction, equation (1) can be rearranged to compare
absorbances:
(3)
where is the loss of material at the interface. For multi-chiral separation: VS =400μL, VB =1180μL and
VT=3540μL.
S2.b. 2D IR fluorescence-excitation (PLE) spectroscopy
2D IR band-gap PLE spectra were recorded using a home-built spectrometer: the sample was excited with a
pulsed Xe-lamp (Edinburgh Instruments, Xe900-xP920), and excitation wavelengths were spectrally selected
with a 300mm grating monochromator (Acton SpectraPro 2355). Emission was collected at 90° and analyzed
using a 150 mm grating spectrograph (Acton SpectraPro 2156) with a liquid nitrogen cooled extended InGaAs
photodiode array detector (Princeton Instruments OMA V:1024/LN-2.2), sensitive up to 2.2 μm. Spectra were
recorded with 5 nm steps in excitation wavelength and an instrumental resolution of 8 nm in excitation and 10
nm in emission wavelength. Appropriate filters were used to eliminate stray light and higher order diffractions
from the spectrometers, and all spectra were corrected for detector and spectrograph sensitivity, filter
transmission, and (temporal and spectral) variations of the excitation light intensity.
S2.c. Resonant Raman spectroscopy
Resonant Raman spectra were excited at multiple laser wavelengths originating either from a Kr+- ion laser,
or from a tunable Rhodamine 6G dye or Ti:Sapphire laser both pumped by an Ar+- ion laser. Spectra were
recorded in backscattering geometry using either a 5-5-5 Princeton Instruments Trivista or a Dilor XY800 triple
Raman spectrometer.
S2.d. Transmission electron microscopy (TEM)
All TEM samples were prepared by dropping ~2 micro liters of SWCNT solutions onto a lacey carbon coated
copper TEM foil. The solution was allowed to wick into a filter paper below and the grid was rinsed with DI
water. All samples were imaged on a FEI Tecnai F30 operated at 300kV.
S2.e. Thermo gravimetric analysis (TGA)
20 ml of solution (ATP separated SWCNT dispersion from bottom phase) is filtered and rinsed to remove
surfactants using 0.2 µm pore diameter membrane (polycarbonate). Resulting 2 mg of the SWCNTs and 2 mg of
raw SWCNTs were used for recording TGA (Thermal Analysis TGA, Q500). The heating rate used was 5°C/min
and heated up to 800°C in air.
S2.f. Atomic force microscope (AFM)
Atomic force microscope (AFM) images were obtained with a Nanoscope IV (Digital
Instruments/Veeco/Bruker Metrology, Inc., SantaBarbara, CA), operating in tapping mode, using an ultrasharp
N-type silicon tip coated with Al on its reflective side for better laser signal quality (APP Nano ACTA, 0.01 –
0.025 ohm/cm, < 10 nm radius, f: 200 – 400 kHz, k: 25 -75 N/m) at a scan-rate of 0.6 Hz and 512 x 512 resolution.
Samples for AFM analysis were prepared with 20 μL of SWNT suspensions spin coated at 5000 RPM onto
roughly 0.4 cm2 freshly cleaved mica surfaces(Ted Pella, Inc., Redding, CA) and rinsed with 2-isopropanol
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
Nano Res.
while spinning to remove the excess of surfactant. Samples were left spinning for 10 min to dry thoroughly.
S2.g. Photoluminescence (PL) images
Individual SWCNT PL images were acquired using an inverted microscope coupled to an EMCCD
(Princeton Instruments ProEM 512). The excitation laser light at 568 nm (Coherent Sapphire LP) was focused on
the back aperture of a 1.49 NA, 100x microscope objective to achieve a wide field imaging spot size of ~20 µm.
Individual SWCNT PL was collected through the same objective and imaged on the EMCCD. 100µL of strongly
diluted separated SWCNT dispersions were mixed with 10µL of tetramethylorthosilicate (TMOS). 10µL of this
TMOS-SWCNT dispersion was sandwiched between acid- and plasma-cleaned quartz slides (EscoOptics). PL
images were obtained with 1 second exposure time.
S3. Additional Figures
Figure S1: PLE maps before and after ATP separation. PLE maps of the starting HiPco 195.2 dispersion, the dispersion after CF, and
the ATP bottom phases before and after CF of the parent solution, showing that the SWCNT chirality distribution remains the same in all
samples. Spectra were calibrated for excitation light intensity and detector sensitivity.
| www.editorialmanager.com/nare/default.asp
Nano Res.
Figure S2: TEM HiPco top and bottom: Less amount of catalyst in the bottom phase.
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
Nano Res.
Figure S3: TGA analysis before and after separation. TGA analysis obtained in air for the starting as-received HiPco 195.2 SWCNTs
and the separated ATP-bottom phase shows that along with bundles, a significant fraction of catalyst particles are removed. For the TGA
analysis, the ATP bottom fraction was first rinsed sufficiently with H2O to remove remaining dextran, PEG and surfactants.
Figure S4: Variation of the SWCNT starting concentration. (a) As-measured absorption spectra of the ATP-separated bottom phase
for 3 different starting concentrations of the parent dispersion. (b) Resonant Raman spectra excited at 785nm, showing no presence of a
(10,2)-bundle Raman vibration in these bottom fractions (dashed red line indicates the expected position). Only a rescaling of the overall
intensity is observed, as the concentration of isolated SWCNTs decreases for a smaller SWCNT starting concentration.
| www.editorialmanager.com/nare/default.asp
Nano Res.
Figure S5: Variation of Dextran polymer molecular weight. (a-b),Volumetric diluted absorption spectra of the ATP-separated bottom
(a) and top (b) phases for different dextran MW at a fixed DOC concentration of 0.088%wt/V. (c-d) Resonant Raman spectra excited
at 785nm, showing the presence of a (10,2)bundle RBM only visible at this wavelength when bundles are present for top and bottom
phases separated using dextran 9K (nominal 9-11K) (c) and dextran 35K (nominal 35-45K) (d). Notice that as the MW of dextran
decreases, relatively more SWCNTs are separated into the bottom phase (including bundles, as is expected for any particles that are
separable with ATP), however, by slightly lowering the DOC concentration, these bundles can be removed to the top phase or interface.
Figure S6: Variation of SC and SDS concentration. Absorption spectra of bottom (a and d) and top (b and e) phases for SC (a-b) and
SDS (d-e) rescaled according to their volumetric dilutions, combined with RRS spectra (c and f) of the bottom phases to study the
presence of bundles. For SC, multi-chiral separation can be achieved only at higher surfactant concentration, while for lower
concentration the distribution of SWCNTs is skewed toward small diameters. For SDS, even sufficiently high surfactant concentration
does not allow the SWCNTs to distribute into the bottom phase, without the presence of bundles or other impurities.
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
Nano Res.
Figure S7: AFM images of Plasma Torch top and bottom fractions. Significant number of SWCNT bundles in the top phase in
comparison to the bottom phase for both types of dispersions
.
| www.editorialmanager.com/nare/default.asp
Nano Res.
Figure S8:TEM Plasma Torch top and bottom fractions. Less amount of catalyst and carbonaceous materials in the bottom phase than
in the top
www.theNanoResearch.com∣www.Springer.com/journal/12274 | Nano Research
Nano Res.
Figure S9: PLE of separated ATP bottom fractions. Multichiral separation from different CNT sources resulted in highly resolved
PLE maps, even for the larger diameter SWCNTs. Spectra were calibrated for excitation light intensity and detector sensitivity
Figure S10: ATP separation for non-sonicated (stirred) samples. ATP separation of the stirred 195.2 HiPco dispersion using different
DOC concentrations for the non-centrifuged sample (rescaled for volumetric dilution). While after centrifugation (e.g. 2h at 40000g or
4h at 16000g) the normal 0.088%wt/V of DOC is sufficient, before centrifugation 0.059%wt/V yields identical separation. We believe
the lower DOC concentration is necessary because of a much higher bundle content (relative to the actual DOC concentration) in the
parent suspension.
400 600 800 1000 1200 1400 1600
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Ab
s (
1 m
m)
Wavelength (nm)
0.059% - no bundles
0.066%
0.073%
0.082%
0.088%
+ CF = 0.088%
Stirred 195.2 HiPco SWCNTs
| www.editorialmanager.com/nare/default.asp
Nano Res.
Figure S11: AFM images of HiPco top and bottom fractions after stirring. Significant number of SWCNT bundles in the top phase in
comparison to the bottom phase and relatively long SWCNTs in the bottom phase can be observed.
Address correspondence to J. G. Duque, [email protected]